* Re: sysv ipc msg functions
2005-04-19 14:42 sysv ipc msg functions Greg Weeks
@ 2005-04-19 14:35 ` Thiemo Seufer
2005-04-19 15:23 ` Daniel Jacobowitz
2005-04-19 14:51 ` sjhill
2005-04-19 15:05 ` Daniel Jacobowitz
2 siblings, 1 reply; 7+ messages in thread
From: Thiemo Seufer @ 2005-04-19 14:35 UTC (permalink / raw)
To: Greg Weeks; +Cc: linux-mips
Greg Weeks wrote:
> I needed this glibc patch to get the sysv ipc msgctl functions to work
> correctly. This looks a bit hackish to me, so I wanted to run it past
> everybody here before filing it with glibc.
The Debian glibc has a similiar patch, see
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=200215&archive=yes
for a discussion.
Thiemo
^ permalink raw reply [flat|nested] 7+ messages in thread
* sysv ipc msg functions
@ 2005-04-19 14:42 Greg Weeks
2005-04-19 14:35 ` Thiemo Seufer
` (2 more replies)
0 siblings, 3 replies; 7+ messages in thread
From: Greg Weeks @ 2005-04-19 14:42 UTC (permalink / raw)
To: linux-mips
[-- Attachment #1: Type: text/plain, Size: 199 bytes --]
I needed this glibc patch to get the sysv ipc msgctl functions to work
correctly. This looks a bit hackish to me, so I wanted to run it past
everybody here before filing it with glibc.
Greg Weeks
[-- Attachment #2: glibc-2.3.3-timesys-mips-msq.h.patch --]
[-- Type: text/x-patch, Size: 1231 bytes --]
--- glibc-2.3.3-200407050320/sysdeps/unix/sysv/linux/mips/bits/msq.h.orig 2005-04-19 09:04:15.000000000 -0400
+++ glibc-2.3.3-200407050320/sysdeps/unix/sysv/linux/mips/bits/msq.h 2005-04-19 09:22:21.000000000 -0400
@@ -38,9 +38,27 @@ typedef unsigned long int msglen_t;
struct msqid_ds
{
struct ipc_perm msg_perm; /* structure describing operation permission */
+#if __WORDSIZE == 32 && defined(__MIPSEB__)
+ unsigned long __unused3;
+#endif
__time_t msg_stime; /* time of last msgsnd command */
+#if __WORDSIZE == 32 && defined(__MIPSEL__)
+ unsigned long __unused3;
+#endif
+#if __WORDSIZE == 32 && defined(__MIPSEB__)
+ unsigned long __unused4;
+#endif
__time_t msg_rtime; /* time of last msgrcv command */
+#if __WORDSIZE == 32 && defined(__MIPSEL__)
+ unsigned long __unused4;
+#endif
+#if __WORDSIZE == 32 && defined(__MIPSEB__)
+ unsigned long __unused5;
+#endif
__time_t msg_ctime; /* time of last change */
+#if __WORDSIZE == 32 && defined(__MIPSEL__)
+ unsigned long __unused5;
+#endif
unsigned long int __msg_cbytes; /* current number of bytes on queue */
msgqnum_t msg_qnum; /* number of messages currently on queue */
msglen_t msg_qbytes; /* max number of bytes allowed on queue */
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: sysv ipc msg functions
2005-04-19 14:42 sysv ipc msg functions Greg Weeks
2005-04-19 14:35 ` Thiemo Seufer
@ 2005-04-19 14:51 ` sjhill
2005-04-19 14:54 ` Greg Weeks
2005-04-19 15:05 ` Daniel Jacobowitz
2 siblings, 1 reply; 7+ messages in thread
From: sjhill @ 2005-04-19 14:51 UTC (permalink / raw)
To: Greg Weeks; +Cc: linux-mips
[ Charset ISO-8859-1 unsupported, converting... ]
> I needed this glibc patch to get the sysv ipc msgctl functions to work
> correctly. This looks a bit hackish to me, so I wanted to run it past
> everybody here before filing it with glibc.
>
Perhaps is ignorance on my part, but I thought the compiler would
handle the endianness with regards to the structure members. Did
you have problems with big and little endian such that you had to
do all of the ugly #ifdef'ing?
-Steve
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: sysv ipc msg functions
2005-04-19 14:51 ` sjhill
@ 2005-04-19 14:54 ` Greg Weeks
0 siblings, 0 replies; 7+ messages in thread
From: Greg Weeks @ 2005-04-19 14:54 UTC (permalink / raw)
To: sjhill; +Cc: linux-mips
sjhill@realitydiluted.com wrote:
>[ Charset ISO-8859-1 unsupported, converting... ]
>
>
>>I needed this glibc patch to get the sysv ipc msgctl functions to work
>>correctly. This looks a bit hackish to me, so I wanted to run it past
>>everybody here before filing it with glibc.
>>
>>
>>
>Perhaps is ignorance on my part, but I thought the compiler would
>handle the endianness with regards to the structure members. Did
>you have problems with big and little endian such that you had to
>do all of the ugly #ifdef'ing?
>
>
yes. Take a look at the kernel structure it is mapping to.
in
asm-mips/msgbuf.h
struct msqid64_ds
Greg Weeks
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: sysv ipc msg functions
2005-04-19 14:42 sysv ipc msg functions Greg Weeks
2005-04-19 14:35 ` Thiemo Seufer
2005-04-19 14:51 ` sjhill
@ 2005-04-19 15:05 ` Daniel Jacobowitz
2005-04-19 15:52 ` Greg Weeks
2 siblings, 1 reply; 7+ messages in thread
From: Daniel Jacobowitz @ 2005-04-19 15:05 UTC (permalink / raw)
To: Greg Weeks; +Cc: linux-mips
On Tue, Apr 19, 2005 at 10:42:24AM -0400, Greg Weeks wrote:
> I needed this glibc patch to get the sysv ipc msgctl functions to work
> correctly. This looks a bit hackish to me, so I wanted to run it past
> everybody here before filing it with glibc.
What's your configuration? Big or little endian, userland ABI, kernel
ABI. Glibc version. Kernel version. What specific things don't work.
Not even enough information here to make a guess.
You're updating the userspace msqid_ds to match the kernel's
msqid64_ds. They're not normally the same type. Rather, see
<linux/msg.h> for the type o32 generally uses.
--
Daniel Jacobowitz
CodeSourcery, LLC
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: sysv ipc msg functions
2005-04-19 14:35 ` Thiemo Seufer
@ 2005-04-19 15:23 ` Daniel Jacobowitz
0 siblings, 0 replies; 7+ messages in thread
From: Daniel Jacobowitz @ 2005-04-19 15:23 UTC (permalink / raw)
To: Thiemo Seufer; +Cc: Greg Weeks, linux-mips
On Tue, Apr 19, 2005 at 04:35:43PM +0200, Thiemo Seufer wrote:
> Greg Weeks wrote:
> > I needed this glibc patch to get the sysv ipc msgctl functions to work
> > correctly. This looks a bit hackish to me, so I wanted to run it past
> > everybody here before filing it with glibc.
>
> The Debian glibc has a similiar patch, see
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=200215&archive=yes
> for a discussion.
Last thing I see there was:
Okay, I suggest you send this patch to Uli for libc and I'll prepare a
patch for the kernel, will post here later.
Anything ever come of that?
--
Daniel Jacobowitz
CodeSourcery, LLC
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: sysv ipc msg functions
2005-04-19 15:05 ` Daniel Jacobowitz
@ 2005-04-19 15:52 ` Greg Weeks
0 siblings, 0 replies; 7+ messages in thread
From: Greg Weeks @ 2005-04-19 15:52 UTC (permalink / raw)
To: Daniel Jacobowitz; +Cc: linux-mips
Daniel Jacobowitz wrote:
>On Tue, Apr 19, 2005 at 10:42:24AM -0400, Greg Weeks wrote:
>
>
>>I needed this glibc patch to get the sysv ipc msgctl functions to work
>>correctly. This looks a bit hackish to me, so I wanted to run it past
>>everybody here before filing it with glibc.
>>
>>
>
>What's your configuration? Big or little endian, userland ABI, kernel
>ABI. Glibc version. Kernel version. What specific things don't work.
>Not even enough information here to make a guess.
>
>You're updating the userspace msqid_ds to match the kernel's
>msqid64_ds. They're not normally the same type. Rather, see
><linux/msg.h> for the type o32 generally uses.
>
>
>
glibc-2.3.3-200407050320
2.6.11.7 kernel but it's the issue is the same on the 2.6.12-rc from cvs.
The board is a malta 4kc in LE mode.
If you want to see a failure the LTP msgctl/msgsnd tests fail.
Greg Weeks
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2005-04-19 15:52 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-04-19 14:42 sysv ipc msg functions Greg Weeks
2005-04-19 14:35 ` Thiemo Seufer
2005-04-19 15:23 ` Daniel Jacobowitz
2005-04-19 14:51 ` sjhill
2005-04-19 14:54 ` Greg Weeks
2005-04-19 15:05 ` Daniel Jacobowitz
2005-04-19 15:52 ` Greg Weeks
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox